U.S. patent number 6,030,666 [Application Number 08/831,495] was granted by the patent office on 2000-02-29 for method for microwave plasma substrate heating.
This patent grant is currently assigned to Lam Research Corporation. Invention is credited to David Hodul, James Lam.
United States Patent |
6,030,666 |
Lam , et al. |
February 29, 2000 |
Method for microwave plasma substrate heating
Abstract
A method of microwave heating of a substrate in a plasma
processing chamber wherein a heatup gas is supplied into the
processing chamber, the heatup process gas is energized with
microwave power to heat an exposed surface of the substrate, a
reactant gas is supplied into the processing chamber and the
reactant gas is energized into a plasma gas state to process the
substrate.
Inventors: |
Lam; James (Oakland, CA),
Hodul; David (Oakland, CA) |
Assignee: |
Lam Research Corporation
(Fremont, CA)
|
Family
ID: |
25259192 |
Appl.
No.: |
08/831,495 |
Filed: |
March 31, 1997 |
Current U.S.
Class: |
427/539;
204/192.3; 257/E21.218; 257/E21.252; 427/543; 427/553; 427/575;
427/588 |
Current CPC
Class: |
C23C
16/402 (20130101); C23C 16/46 (20130101); H01L
21/31116 (20130101); H01L 21/3065 (20130101); C23C
16/511 (20130101) |
Current International
Class: |
C23C
16/511 (20060101); C23C 16/40 (20060101); C23C
16/46 (20060101); C23C 16/50 (20060101); H01L
21/02 (20060101); H01L 21/3065 (20060101); H01L
21/311 (20060101); H05H 001/00 () |
Field of
Search: |
;118/722,723MP,723MW
;427/571,539,535,575,543,579,574,573,533,588 |
References Cited
[Referenced By]
U.S. Patent Documents
Foreign Patent Documents
|
|
|
|
|
|
|
63-186874 |
|
Aug 1988 |
|
JP |
|
1-298173 |
|
Dec 1989 |
|
JP |
|
5-86479 |
|
Apr 1993 |
|
JP |
|
7-335570 |
|
Dec 1995 |
|
JP |
|
WO 91/00613 |
|
Jan 1991 |
|
WO |
|
Primary Examiner: Nguyen; Nam
Assistant Examiner: Ver Steeg; Steven H.
Attorney, Agent or Firm: Burns, Doane, Swecker & Mathis,
L.L.P.
Claims
What is claimed is:
1. A method of processing an exposed surface of a substrate in a
plasma processing chamber, comprising steps of:
supplying heatup gas comprising oxygen into a plasma processing
chamber containing a substrate to be processed;
heating the substrate by energizing the heatup gas into a plasma
gas state with microwave power so as to heat an exposed surface of
the substrate with the microwave energized heatup gas, the
substrate being heated such that the exposed surface reaches a
threshold temperature and the substrate being heated without
applying an rf bias to the substrate;
supplying reactant gas into the plasma processing chamber; and
processing the substrate by energizing the reactant gas into a
plasma gas state so as to process the exposed surface of the
substrate with the energized reactant gas.
2. The method of claim 1, wherein the heatup gas further comprises
argon, krypton, xenon, nitrogen, helium, neon and mixtures
thereof.
3. The method of claim 1, wherein during the heating step the
microwave power is applied at a power level of at least 1000
watts.
4. The method of claim 1, wherein during the heating step the
heatup gas is supplied at a rate of at least 50 sccm.
5. The method of claim 1, wherein during the heating step the
microwave power is adjusted to heat the exposed surface to at least
90% of a processing temperature at which the exposed surface is
processed during the processing step.
6. The method of claim 1, wherein the substrate comprises a
semiconductor wafer and the processing step comprises depositing a
dielectric layer on the exposed surface of the wafer.
7. The method of claim 6, wherein the reactant gas includes a
silicon-containng gas and an oxygen-containing gas.
8. The method of claim 7, wherein the silicon-containing gas
comprises SiH.sub.4 or SiF.sub.4 and the oxygen-containing gas
comprises O.sub.2.
9. The method of claim 1, wherein during the heating step the
exposed surface is heated without applying a chucking force or
backside cooling to the substrate.
10. The method of claim 9, wherein the substrate is supported on an
electrostatic chuck during the heating and processing steps.
11. The method of claim 1, wherein the plasma processing chamber
comprises an ECR tool having a plasma formation chamber adjacent a
reaction chamber, the substrate being supported in the reaction
chamber.
12. The method of claim 1, wherein microwave power is supplied to
the plasma processing chamber during the heating and processing
steps.
13. The method of claim 1, wherein the threshold temperature is at
least 200.degree. C.
14. The method of claim 1, wherein the heating step is performed in
less than 60 seconds and the exposed surface is sputtered during
the heating step at a sputtering rate of less than 500
.ANG./min.
15. The method of claim 14, wherein the sputtering rate is less
than 100 .ANG./min.
16. The method of claim 14, wherein the sputtering rate is less
than 50 .ANG./min.
17. The method of claim 1, wherein the substrate is supported on an
electrostatic chuck during the heating and processing steps, the
heating step being performed without applying a clamping force to
the substrate and without backside cooling of the substrate, and
the processing step being performed while applying a clamping force
and backside cooling to the substrate.
18. The method of claim 1, wherin the exposed surface is processed
during the processing step with a highly density plasma.
Description
FIELD OF THE INVENTION
The present invention relates generally to a method for heating a
substrate with microwave energy. More specifically, the present
invention relates to a method for heating a substrate with
microwave energy in preparation for chemical vapor deposition.
BACKGROUND OF THE INVENTION
Chemical vapor deposition (CVD) is a process for forming a material
layer on a substrate by the reaction of gas phase reactants at or
near a substrate surface. Vacuum processing chambers are generally
used for CVD and etching of materials on substrates by supplying
process gas to the vacuum chamber and applying a radio frequency
(rf) field to the gas. Examples of parallel plate, transformer
coupled plasma (TCP.TM., also called inductively coupled plasma or
ICP), and electron-cyclotron resonance (ECR) reactors are disclosed
in commonly owned U.S. Pat. Nos. 4,340,462; 4,948,458; and
5,200,232. The substrates are held in place within the vacuum
chamber during processing by substrate holders. Conventional
substrate holders include mechanical clamps and electrostatic
clamps (ESC). Examples of mechanical clamps and ESC substrate
holders are provided in commonly owned U.S. Pat. No. 5,262,029 and
commonly owned U.S. application Ser. No. 08/401,524 filed on Mar.
10, 1995. Substrate holders in the form of an electrode can supply
rf power into the chamber, as disclosed in U.S. Pat. No.
4,579,618.
Plasma processing systems wherein an antenna coupled to an rf
source energizes gas into a plasma state within a processing
chamber are disclosed in U.S. Pat. Nos. 4,948,458; 5,198,718;
5,241,245; 5,304,279; and 5,401,350. In such systems, the antenna
is located outside the processing chamber and the rf energy is
supplied into the chamber through a dielectric window. Such
processing systems can be used for a variety of semiconductor
processing applications such as etching, deposition, resist
stripping, etc.
Thermal CVD processes typically rely on heating of the substrate
surface in order to promote the reaction(s) which result in
formation of compound(s) on the substrate surface. Conventionally,
the substrate surface is heated by applying an rf bias voltage to
the substrate. The use of the rf bias voltage to heat a substrate
to promote the reactions generates a DC bias in the plasma near the
surface of the substrate. This DC bias accelerates ions toward the
substrate surface, and the ensuing collision of the ions transfers
energy to the substrate, thereby heating the substrate.
A problem with this conventional heating of the substrate surface
with an rf bias voltage is that it causes sputtering of the
material on the substrate due to ion bombardment. This can cause
damage to the substrate surface. Various gases have been used as
the reacting gases in an attempt to lessen the sputtering while
heating the substrate with the rf bias voltage. The problem of
sputtering remains, however.
Other methods of heating the substrate surface have also been
proposed, including heating lamps, resistor heaters, and other
in-situ substrate heating methods. These methods are limited by
slow response time and process variability due to previous
processing of the substrate or the condition of the substrate.
There is thus a need for a method for minimizing sputtering during
heating of a substrate surface for processing such as CVD
processing.
SUMMARY OF THE INVENTION
Accordingly, it is an object of the present invention to provide a
process for minimizing sputtering while heating a substrate surface
as an initial step in plasma processing such as CVD processing.
This and other objects are met by a method of microwave heating of
a substrate in a plasma processing chamber wherein a heatup gas is
supplied into the processing chamber, the heatup process gas is
energized with microwave power to heat an exposed surface of the
substrate, a reactant gas is supplied into the processing chamber
and the reactant gas is energized into a plasma gas state to
process the substrate.
BRIEF DESCRIPTION OF THE DRAWINGS
The foregoing, and other objects, features, and advantages of the
present invention will be more readily understood upon reading the
following detailed description in conjunction with the drawings in
which:
FIG. 1 illustrates graphically a temperature profile of a substrate
surface during a CVD process;
FIG. 2 illustrates graphically heatup times for a substrate surface
applying microwave power to various reactant gases;
FIGS. 3a and 3b illustrate graphically a heatup time for a
substrate surface applying microwave power to argon;
FIG. 4a and 4b illustrate graphically a heatup time for a substrate
surface applying microwave power to oxygen;
FIG. 5 illustrates graphically a heatup time for a substrate
surface applying microwave power to helium;
FIG. 6 illustrates graphically a temperature profile of a substrate
surface during a CVD process applying microwave power in a heatup
step;
FIG. 7 illustrates graphically a temperature profile of a substrate
surface during a conventional CVD process wherein microwave power
and rf bias power are used during a heatup step;
FIG. 8 illustrates graphically a temperature profile of a substrate
surface during a CVD process with 8 torr of helium backside
pressure; and
FIG. 9 illustrates graphically a temperature profile of a substrate
surface during a CVD process with 10 torr of helium backside
pressure.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
According to the present invention, a method is provided for
minimizing sputtering while heating a substrate surface (such as a
semiconductor substrate, flat panel display, etc.) as a step in
processing of the substrate. For example, the substrate surface can
be heated in preparation for CVD processing, etch processing, etc.,
by exciting a heatup gas in a plasma processing chamber. The heatup
gas can be composed of one or more ionizable reacting or
non-reacting gases including, for example, oxygen, argon, silane,
silicon-tetraflouride, helium, neon, krypton, xenon, nitrogen or
mixtures thereof. The plasma can be provided by any suitable source
such as an electron cyclotron resonance (ECR) tool, a parallel
plate reactor, an inductively coupled reactor, a transformer,
coupled reactor, a helicon, a helical resonator or the like.
According to a first embodiment wherein the present invention is
used in conjunction with a CVD process, the substrate surface is
heated as an initial step in high density plasma (e.g. 10.sup.11
-10.sup.12 ions/cm.sup.3) processing by applying microwave power to
a heatup gas in a plasma processing chamber of a DSM.TM. 9900 or an
EPIC.TM. tool manufactured by Lam Research Incorporated, the
Assignee of the present application. FIG. 1 illustrates graphically
a temperature profile of a substrate surface during CVD of a
SiO.sub.x film on the substrate. The CVD processing begins by
energizing a heatup gas with microwave power to heat the substrate
surface. FIG. 1 illustrates a heating profile which results by
energizing oxygen gas with microwave power to heat the substrate
surface to approximately 90% of its processing temperature, in this
example 270.degree. C. As shown in FIG. 1, this initial heating
step takes approximately 38 seconds. Next, suitable reactant gases
are supplied and actual processing of the substrate begins, with
switching on of the rf bias voltage supplied by the substrate
support, causing a small jump in substrate temperature to
approximately 300.degree. C. The temperature of the substrate
surface gradually decreases during the CVD processing to about
270.degree. C., due to the ESC voltage and He backside temperature
control. At around 130 seconds, the processing ends and dechucking
begins, which causes the temperature of the substrate surface to
decrease. When the dechucking step ends, the temperature rises
again.
According to the first embodiment of the present invention, by
eliminating rf bias power supplied by the substrate support from
the initial step of heating the substrate surface, the DC bias in
the plasma is greatly reduced. Thus, sputtering of the substrate
surface to be processed is greatly reduced. The energy used to heat
the substrate is preferably supplied only by the microwave energy.
Heating rates can be adjusted by controlling the microwave power
density in the plasma. The heating method according to the first
embodiment of the present invention is thus independent of many
substrate variables, such as substrate backside roughness and
previous processing.
In the following figures, application of a chucking force to the
substrate and supply of helium back pressure between the substrate
holder and the substrate are omitted during the initial heatup step
to minimize the heat loss to the substrate support and reduce the
time of the heatup step. The processes are carried out in a LAM
DSM.TM. tool and the substrate surface temperatures shown in the
figures represent measurements using a convection probe and/or
phosphor dot on the backside of the substrate. Sputter rates of
these heatups were recorded on both oxide coated and TiN coated
substrates by running the heatup and measuring the before and after
thickness of the oxide and TiN coatings. Compared to a sputter rate
on the order of 1000 .ANG./min when using rf bias power for heating
a substrate coated with SiO.sub.2, the microwave heating according
to the invention can reduce the sputter rate well over 50%,
preferably over 75%, and more preferably over 90% to around 20
.ANG./min.
FIG. 2 illustrates graphically heatup times for a substrate surface
applying microwave power to various heatup gases. In FIG. 2, the
heatup times for oxygen, argon, and helium heatup gases are
represented. The microwave power applied to heat up these gases is
constant at 1800 watts. The line with superimposed diamonds
represents oxygen flowing at 200 scam, the line with superimposed
boxes represents argon flowing at 200 sccm, and the line with
superimposed triangles represents helium flowing at 175 sccm. As
can be seen from FIG. 2, under the described conditions, helium
heats up the substrate fastest, heating to 270.degree. C. in
approximately 24 seconds, followed by argon which heats the
substrate to 270.degree. C. in approximately 30 seconds, and then
oxygen, which heats the substrate to 270.degree. C. in
approximately 40 seconds.
FIGS. 3a and 3b illustrate graphically heatup times for a substrate
surface applying microwave power to argon. In FIG. 3a, argon is
energized with a constant amount of microwave power set at 1800
watts. The line with superimposed diamonds represents argon flowing
at 100 sccm, and the line with superimposed boxes represents argon
flowing at 200 sccm. As can be seen from FIG. 3a, argon flowing at
100 sccm heats to 270.degree. C. approximately 2 seconds faster
than argon flowing at 200 sccm.
In FIG. 3b, the flow rate of argon is constant at 100 sccm. The
line with superimposed diamonds represents argon energized with
1400 watts of microwave power, and the line with superimposed boxes
represents argon energized with 1800 watts of microwave power. As
can be seen from FIG. 3b, argon energized with 1800 watts heats to
270.degree. C. approximately 10 seconds faster than argon energized
with 1400 watts.
FIG. 4a and 4b illustrate graphically a heatup time for a substrate
surface applying microwave power to oxygen. In FIG. 4a, the oxygen
is heated with a constant amount of microwave power set at 1800
watts. The line with superimposed triangles represents oxygen
flowing at 300 sccm, the line with superimposed diamonds represents
oxygen flowing at 200 sccm, and the line with superimposed boxes
represents oxygen flowing at 100 sccm. As can be seen from FIG. 4a,
oxygen flowing at 100 sccm heats to 270.degree. C. the fastest, in
approximately 35 seconds, followed by oxygen flowing at 200 sccm,
and then oxygen flowing at 300 sccm.
In FIG. 4b, the flow rate of oxygen is constant at 100 sccm. The
line with superimposed diamonds represents oxygen energized with
1000 watts of microwave power, the line with superimposed boxes
represents oxygen energized with 1400 watts of microwave power, and
the line with superimposed triangles represents oxygen energized
with 1800 watts of microwave power. As can be seen from FIG. 4b,
oxygen energized with 1800 watts heats up to 270.degree. C. the
fastest, in approximately 38 seconds, followed by oxygen energized
at 1400 watts, and then oxygen energized at 1000 watts, which takes
approximately 140 seconds.
As can be seen from FIGS. 3a,b and 4a,b, the oxygen and argon
plasmas are very stable at various flow rates and microwave powers.
In contrast, the helium plasma is unstable at low flow rates. For
instance, at 100 sccm and 140 sccm He, the He plasma is unstable or
would not light. However, flow rates of 175 scam He provide a
stable plasma, especially at higher microwave power levels. This
can be seen from FIG. 5 which illustrates graphically the heatup
time for a substrate surface applying microwave power to helium. In
FIG. 5, the flow rate of helium is constant at 175 sccm. The line
with superimposed diamonds represents helium energized with 1800
watts of microwave power, and the line with superimposed boxes
represents helium energized with 1400 watts of microwave power. As
can be seen from FIG. 5, there is little difference between the
heatup times to 270.degree. C. with microwave power at 1400 watts
and 1800 watts, although the microwave power at 1800 watts heats
the helium to 270.degree. C. approximately a second faster. Thus,
while He may be difficult to ionize and remove from the reaction
chamber with a turbo pump, it otherwise is a good candidate for
substrate heating.
Auxiliary gases such as argon or oxygen can be added to the helium
to help ignite the He plasma, but the subsequent removal of such
auxiliary gases may extinguish or destabilize the He plasma.
Although not shown, a similar result is achieved for neon. While
most neon plasmas ignite readily, it is, like helium, unstable at
low flow rates. Further, at 100 sccm Ne the reflected power is
about one-half of the applied power, whereas at 200 sccm, the
microwaves tuned perfectly. The heatup time is about 90 seconds for
neon flowing at 100 sccm.
Another candidate, nitrogen, produces less satisfactory results,
with a heatup time of about 47 seconds under the same conditions.
Also, nitrogen might have an effect on the material that is being
heated up and produce nitrides, like oxides when an oxygen plasma
is used.
According to the first embodiment of the present invention, heating
a substrate surface using only microwave energy is achievable with
comparable times to those conditions using an rf bias voltage to
heat the substrate surface, if the substrate is not chucked during
the heatup step. Argon may be preferable to He or oxygen for
heating up the substrate surface since argon heats the substrate to
the desired temperature in about 25 seconds compared to 30 seconds
for oxygen. Also, if oxygen is used, a side effect of an oxygen
plasma as a descumer should be considered. The difference in
heating rate for oxygen and argon appears to be due to the mass of
the ions in the plasma.
There is a small correlation between gas flows and heat up times.
This change is probably due to a higher ion flux, and higher ion
density in the plasma. There seems to be little difference in the
heatup rates using argon, but the difference is noticeable with
oxygen. Microwave power is the major factor in the amount of energy
applied to the substrate. With increasing power, the heatup time is
reduced, possibly due to higher ion density from the increased
power.
A heated chuck can help heat the substrate surface, even without
chucking the wafer. For a helium plasma, using a heated chuck
results in faster heatup times, quickly bringing the substrate
surface temperature to about 400.degree. C. in about 11 seconds for
a low flow rate and low microwave power. However, use of a heated
chuck presents a problem of cooling the substrate during the CVD
process, the effective cooling rate of the backside helium being
greatly lessened.
The heatup process according to the first embodiment of the present
invention greatly minimizes sputtering of the exposed substrate
surface. In the case of silicon dioxide coated silicon wafers,
compared to the conventional heatup using rf bias energy which
substantially abates SiO.sub.2 on the order of 1000 .ANG./min,
heating the substrate with microwave energy according to the
invention provides minimal ablation of the SiO.sub.2 at a rate of
approximately 20 .ANG./min. This is a new and unexpected result
achieved according to the process of the invention. There is little
effect in sputter rate due to ion mass, and the bias induced by the
microwave plasma is very small, under the threshold of sputtering.
According to the invention, the uniformity of the substrate
temperature is related only to the uniformity of the plasma. Other
factors that typically affect the temperature uniformity in the
case of conventional substrate heating such as the chuck, rf bias,
and backside helium, are not factors in substrate surface
temperature uniformity when only using the microwave power to
excite the plasma.
Regarding tuning of the microwaves, there is a lag time during the
changeover of the gases during the process. For example, when
reactant gases such as silane and other gases are supplied to the
reaction chamber to start the processing of the substrate, and the
heatup gases are switched off, there is a dramatic change in the
impedance of the plasma and a large percentage of redirected power
during this tune time. To minimize the disturbance, the gases that
are used to heat up the substrate are preferably compatible with or
utilize one or more of the reactant gases used in the CVD
process.
According to exemplary embodiments, Table 1 shows heatup and
sputtering rates for various heatup gases which can be used for
heating a substrate prior to depositing a layer of material such as
silicon dioxide.
TABLE 1 ______________________________________ Minimum Heatup Time:
0.sub.2 : 35 sec. 5" and 8" to 270.degree. C. 60 sec. to 370 8" Ar:
27 sec. 5" and 8" 40 sec. to 350 8" He: 22 sec. 5" Ne: 90 sec. 5"*
N.sub.2 : 47 sec. 5"* 18 sec. Baseline w/rf: *Plasma not stable
Sputtering rate of SiO.sub.2 0.sub.2 : 10.ANG./min avg 16.ANG./min
max 8" 21.ANG./min avg 42.ANG./min max 5" Baseline w/rf:
1048.ANG./min avg 1379.ANG./min max 8"
______________________________________
Processing conditions typically depend on the CVD process and on
the type and/or size of substrate. For instance, during CVD
processing the rf bias power can be scaled to the size of the
substrate. The microwave heatup process can produce an initial
heatup of the substrate surface to 270.degree. C., with the
substrate surface staying within 20.degree. of 270.degree. C.
throughout the deposition process.
Minor modifications may be needed for smooth substrates and
possibly patterned substrates. The process is effective for both
convection probes and infrared ("IR") probes. Preferably,
substrates to be run should be tested with a phosphor dot substrate
if using the convection probe, to insure that there are no
temperature anomalies. The "IR" probes and the phospor dot
substrates are direct measures of temperatures while the convection
probe is a secondary measurement.
In the following figures, substrate surface temperatures measured
with a convection probe and a phosphor dot on the backside of the
substrate are represented.
FIG. 6 illustrates graphically a temperature profile of a substrate
surface during CVD processing applying microwave power to heat the
substrate surface. In FIG. 6, the line with superimposed diamonds
represents results using a convection probe, and the line with
superimposed boxes represent results using a phosphor dot.
FIG. 7 illustrates graphically a temperature profile of a substrate
surface during CVD processing applying conventional rf bias power
from the substrate support to heat the substrate surface. In FIG.
7, the line with superimposed diamonds represents results using a
convection probe, and the line with superimposed boxes represent as
results using a phosphor dot.
There are distinct differences between FIGS. 6 and 7. The most
obvious difference is the temperature overshoot. In FIG. 7, the
initial heatup step heats the substrate surface to 350.degree. C.,
while in FIG. 6, the initial heatup step heats the substrate
surface to 330.degree. C.
The convection probe that is used to monitor the substrate surface
temperature responds very slowly to the actual substrate surface
temperature. Thus, while the temperature profile appears to read
the 270.degree. C. temperature much quicker, this is actually due
to the higher overshoot of the actual substrate temperature.
Varying the minimum backside helium or helium pressure does not
make a significant difference in the response of the convection
probe. This can be seen from comparing FIGS. 8 and 9.
FIG. 8 illustrates graphically a temperature profile of a substrate
surface during CVD processing with 8 torr helium backside pressure.
In FIG. 8, the line with superimposed diamonds represents results
using a convection probe, and the line with superimposed boxes
represents results using a phosphor dot.
FIG. 9 illustrates graphically a temperature profile of a substrate
surface during CVD processing with 10 torr helium backside
pressure. In FIG. 9, the line with superimposed diamonds represents
results using a convection probe, and the line with superimposed
boxes represents results using a phosphor dot.
The microwave convection probe data suggests a faster response, but
this is, again, due to the higher substrate temperature. Helium
backside pressure does not play a significant part in changing the
response time of the convection probe at typical helium backside
pressures between approximately 5 and 10 torr.
In the embodiment described above, heating of a substrate surface
as an initial step in preparation for CVD processing is performed
by the application of microwave power. In an alternate embodimaent,
the microwave power can be supplemented with low levels of if power
to assist in the heating.
The above-described exemplary embodiments are intended to be
illustrative in all respects, rather than restrictive, of the
present invention. Thus, the present invention is capable of many
variations in detailed implementation that can be derived from the
description contained herein by a person skilled in the art. All
such variations and modifications are considered to be withi the
scope and spirt of the present invention as defined by the
following claims.
* * * * *